beta kernel final notebook

This is an R Markdown Notebook. When you execute code within the notebook, the results appear beneath the code.

Try executing this chunk by clicking the Run button within the chunk or by placing your cursor inside it and pressing Ctrl+Shift+Enter.

Add a new chunk by clicking the Insert Chunk button on the toolbar or by pressing Ctrl+Alt+I.

When you save the notebook, an HTML file containing the code and output will be saved alongside it (click the Preview button or press Ctrl+Shift+K to preview the HTML file).

The preview shows you a rendered HTML copy of the contents of the editor. Consequently, unlike Knit, Preview does not run any R code chunks. Instead, the output of the chunk when it was last run in the editor is displayed.

setwd("D:/R program")

The working directory was changed to D:/R program inside a notebook chunk. The working directory will be reset when the chunk is finished running. Use the knitr root.dir option in the setup chunk to change the working directory for notebook chunks.

#values needed 
K= 1.38064852*(10)^-23 #m2 kg/ s2 K boltzmann constant
mu= 1.126*(10)^-3 #kg/m s dynamic viscosity in 18C
v= 1.099*(10)^-6 #m2/s kinematic viscosity in 18C
Reh_calc= 2.3E-6 #in m radius Ehux
Reh_naked= 1.8E-6 #in m radius Ehux
Rehv= 90*(10)^-9 #in m radius virus
Temp = 18+273.15 #temp in kelvin, here assuming 18C
Den_OcM = 1.05 #g/cm3 density organic cell matter
Den_CH2O= 1.025 #g/cm3 density seawater at 18C
hostnum <- (10)^3
virnum <- (10)^4
require (ggplot2)
require(plotly)
require(grid)
require(ggthemes)
require (dplyr)
require(plyr)
source ("theme_Publication.R")
source("resizewin.R")
resize.win(12,9)

Error in dev.off() : cannot shut down device 1 (the null device)

for Brownian motion

#Brownian motion (BM)
#1. make a data frame
BM <- data.frame (group= c("naked", "calcified"), rad= c(1.8E-6, 2.3E-6)) 
#2. calculate beta (beta)
BM$beta_s <- (2*(K*(10)^4)*Temp*(((BM$rad+Rehv)*100)^2))/((3*mu*10)*(BM$rad*Rehv*1e4)) #m3/s
BM$beta_d <- BM$beta_s*86400 #to cm3/day
# go back to this later
#3. calculate encounters (E)
BM$E <- BM$beta_d*hostnum
BM$E_HV <- BM$beta_d*virnum*hostnum
BM

Differential settling (DS)

#Differential settling (DS)
#1. read in PIC data
library(readr) #always use readr not baseR
setwd("D:/R program")

The working directory was changed to D:/R program inside a notebook chunk. The working directory will be reset when the chunk is finished running. Use the knitr root.dir option in the setup chunk to change the working directory for notebook chunks.

PIC <- read_csv("Postdoc-R/CSV Files/PIC.csv")

Parsed with column specification:
cols(
  Strain = col_character(),
  Replicate = col_integer(),
  TC = col_double(),
  AC = col_double(),
  Cellcount = col_double()
)

PIC$Strain <- as.factor(PIC$Strain)
PIC$Replicate <- as.factor(PIC$Replicate)
#certain changes in data.table API made calculating inside the list data.table to not work
#2. calculate PIC
PIC$PIC <- PIC$TC-PIC$AC
PIC$PICpercell <- (PIC$PIC/PIC$Cellcount)*(10)^-3#in g
PIC$PICpercellpg <- PIC$PICpercell*1e12
ggplotly(ggplot(data=PIC, aes(x=Strain, y=PICpercellpg)) + geom_boxplot()+geom_point(size=2) +theme_Publication())

#3. calculate density of cells (den)
PIC <- mutate(PIC, group = ifelse(PICpercellpg < 4 , "naked", "calcified"))
ggplotly(ggplot(data=PIC, aes(x=group, y=PICpercellpg)) + geom_boxplot()+geom_point(size=2, aes(color=Strain))+
           theme_Publication())

plotly.js does not (yet) support horizontal legend items 
You can track progress here: 
https://github.com/plotly/plotly.js/issues/53 

PIC <- mutate(PIC, rad = ifelse(group == "naked" ,  1.8E-6,  2.3E-6)) #in m
PIC$volume <- (4/3)*pi*(PIC$rad*100)^3 #in cm3
PIC$Den_cell <- PIC$PICpercell/PIC$volume #g/cm3
PIC$Den_celltotal <- PIC$Den_cell+Den_OcM
ggplotly(ggplot(data=PIC, aes(x=Strain, y=Den_celltotal, color=group)) + geom_boxplot()+geom_point(size=2) 
         +theme_Publication())

plotly.js does not (yet) support horizontal legend items 
You can track progress here: 
https://github.com/plotly/plotly.js/issues/53 

#some strains that are "naked" have PIC<2. I chose to ignore this since in the lm model I do not use
#strain as a factor, rather data is treated as a whole (e.g., no grouping)
#4. calculate sinking velocity of cells
PIC$SinkVel <- ((2*((PIC$rad*100)^2)*(981)*(PIC$Den_celltotal-Den_CH2O))/(9*(mu*10)))*864 #meter per day
#g is converted to per day, 864 is the one that converts cm/s to m/day
#plot sinking velocity vs calcification
ggplot(data=PIC, aes(x=PICpercellpg, y=SinkVel, color=Strain, shape=group)) + geom_point(size=5)+theme_Publication()+
    labs(y = expression("Sinking velocity"~("m"~day^-1)), x = expression("PIC"~cell^-1)) +
    scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=2),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) + annotation_logticks(sides="l")

#5. calculate sinkvel of viruses
Den_virus <- 1.09 #data from Ben D. fresh EhV-207 density. old density of EhV-207 is 1.19
Ehv_SinkVel <- ((2*((Rehv*100)^2)*(981)*(Den_virus-Den_CH2O))/(9*(mu*10)))*864  #equals to 0
#6. calculate beta kernels
PIC$beta_s <- pi*(((PIC$rad+Rehv)*100)^2)*(abs((PIC$SinkVel-Ehv_SinkVel)/864)) #in encounters cm3/s
PIC$beta_d <- PIC$beta_s*86400 #in cm3/day
Sinkvelbeta.plot<- ggplot(data=PIC, aes(x=SinkVel, y=beta_d, color=Strain, shape=group)) + geom_point(size=5)+
  theme_Publication()+
  labs(x = expression("Sinking velocity"~("m"~day^-1)), y = expression(beta~("Encounters" ~ cm^3~day^-1))) +
 scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=2),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) + annotation_logticks(sides="l")
Sinkvelbeta.plot #change ticks

ggplotly(Sinkvelbeta.plot)

geom_GeomLogticks() has yet to be implemented in plotly.
  If you'd like to see this geom implemented,
  Please open an issue with your example code at
  https://github.com/ropensci/plotly/issuesplotly.js does not (yet) support horizontal legend items 
You can track progress here: 
https://github.com/plotly/plotly.js/issues/53 

ggplotly(ggplot(data=PIC, aes(x=Strain, y=SinkVel)) + geom_boxplot()+theme_Publication())

ggplot(data=PIC, aes(x=PICpercellpg, y=beta_d, color=Strain)) + geom_point(size=5)+theme_Publication()+
  labs(y = expression(beta~("Encounters"~cm^3~day^-1)), x = expression("PIC"~cell^-1))  +
scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=2),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) + annotation_logticks(sides="l")

#7. calculate beta and encounters
#beta are in cells cm3/ day then encounters are to cells/cm3 day
PIC$E_DS_HV <- (PIC$beta_d*virnum*hostnum)  #E calculated with Virus and Host (10:1 MOI)
PIC$E_DS_V <- (PIC$beta_d*virnum) #E calculated with Virus
#8. calculate for lith parameters
lithvol <- 3*1e-12 #in cm3, from CJ's paper
PIC$perlith <- PIC$PICpercell/20 #in g, assuming 20 liths attached
PIC$perlithpg <- PIC$perlith*1e12 #in pg
PIC$Denlith <- (PIC$perlith/lithvol) + Den_OcM #in g/cm3, with organic matter attached
rad_lith <- 2E-6 #in m radius
PIC$SinkVel_lith <- ((2*((rad_lith*100)^2)*(981)*(PIC$Denlith-Den_CH2O))/(9*(mu*10)))*864 #meter per day
PIC$beta_s_lith <- pi*(((rad_lith+Rehv)*100)^2)*(abs((PIC$SinkVel_lith-Ehv_SinkVel)/864)) #in encounters cm3/s
PIC$beta_d_lith <- PIC$beta_s_lith*86400 #in cm3/day
ggplot(data=PIC, aes(x=perlithpg, y=SinkVel_lith, color=Strain, shape=group)) + geom_point(size=5)+theme_Publication()+
  labs(y = expression("Sinking velocity"~("m"~day^-1)), x = expression("PIC"~lith^-1)) 

PIC$Elith_DS_HV <- (PIC$beta_d_lith*virnum*hostnum)  #E calculated with Virus and Host (10:1 MOI)
PIC$Elith_DS_V <- (PIC$beta_d_lith*virnum) #E calculated with Virus
require (dplyr)
PIC$group2 <- case_when(
  PIC$PICpercellpg <2  ~ "naked_bouyant",
  PIC$PICpercellpg >2 & PIC$PICpercellpg < 4 ~ "naked/calcified uncertain",
  PIC$PICpercellpg >4 & PIC$PICpercellpg < 10 ~ "moderately calcified",
  PIC$PICpercellpg >10 ~ "strongly calcified", 
  TRUE ~ as.character(PIC$PICpercellpg)
)
breaks <- 10^(-10:10)
ggplot(data=PIC, aes(x=SinkVel, y=E_DS_HV, color=Strain, shape=group)) + geom_point(size=5)+theme_Publication()+
 scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=2),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) + annotation_logticks(sides="l")

ggplot(data=PIC, aes(x=SinkVel, y=E_DS_V, color=Strain, shape=group)) + geom_point(size=5)+theme_Publication()+
  scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=2),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) + annotation_logticks(sides="l")

ggplot(data=PIC, aes(x=SinkVel_lith, y=Elith_DS_V, color=Strain, shape=group)) + geom_point(size=5) + 
  theme_Publication() + scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=3),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) + annotation_logticks(sides="l")

summary_DS <- ddply(PIC, .(Strain), summarize,  PICpercellpg=mean(PICpercellpg), perlithpg = mean(perlithpg), 
                    Den_celltotal = mean (Den_celltotal),
                    SinkVel=mean(SinkVel),beta_d=mean(beta_d), E_DS_V= mean(E_DS_V), E_DS_HV=mean(E_DS_HV),
                    SinkVel_lith=mean (SinkVel_lith), beta_d_lith=mean (beta_d_lith), Elith_DS_HV=mean (Elith_DS_HV),
                    Elith_DS_V=mean (Elith_DS_V))
summary_DS_bygroup <- ddply(PIC, .(group2), summarize,  PICpercellpg=mean(PICpercellpg), perlithpg = mean(perlithpg), 
                            Den_celltotal = mean (Den_celltotal),
                            SinkVel=mean(SinkVel),beta_d=mean(beta_d), E_DS_V= mean(E_DS_V), E_DS_HV=mean(E_DS_HV), 
                            SinkVel_lith=mean (SinkVel_lith), beta_d_lith=mean (beta_d_lith), 
                            Elith_DS_HV=mean (Elith_DS_HV), Elith_DS_V=mean (Elith_DS_V))
summary_DS

summary_DS_bygroup

setwd("D:/R program")

The working directory was changed to D:/R program inside a notebook chunk. The working directory will be reset when the chunk is finished running. Use the knitr root.dir option in the setup chunk to change the working directory for notebook chunks.

require(openxlsx)
write.xlsx(summary_DS, file = "Postdoc-R/Exported Tables/summary_DS.xlsx")
write.xlsx(summary_DS_bygroup, file = "Postdoc-R/Exported Tables/summary_DS_bygroup.xlsx")

#9. regression of PIC and sinkvel of cells and liths
#a. for cells
PIC_reg <- lm(SinkVel~PICpercellpg, data=PIC) #essentially perfect fit: summary may be unreliable haha
summary(PIC_reg)

Call:
lm(formula = SinkVel ~ PICpercellpg, data = PIC)

Residuals:
       Min         1Q     Median         3Q        Max 
-0.0115036 -0.0029329  0.0004329  0.0021756  0.0113796 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept)  0.0180085  0.0009620   18.72   <2e-16 ***
PICpercellpg 0.0177476  0.0001254  141.58   <2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.005175 on 44 degrees of freedom
Multiple R-squared:  0.9978,    Adjusted R-squared:  0.9978 
F-statistic: 2.004e+04 on 1 and 44 DF,  p-value: < 2.2e-16

plot(residuals.lm(PIC_reg))
layout(matrix(1:4,2,2))

plot(PIC_reg)

coef(PIC_reg)

 (Intercept) PICpercellpg 
  0.01800852   0.01774764 

# coef(PIC_reg)
#(Intercept) PICpercellpg 
#0.01800852   0.01774764 
cor(PIC$PICpercellpg, PIC$SinkVel)

[1] 0.9989042

#cor = 0.9989042
beta_reg <- lm(beta_d~PICpercellpg, data=PIC)
plot(residuals.lm(beta_reg))
coef(beta_reg)

 (Intercept) PICpercellpg 
1.639233e-07 3.243054e-07 

#coef(beta_reg)
# (Intercept) PICpercellpg 
#1.639233e-07 3.243054e-07
E_DS_HV_reg <- lm(E_DS_HV~PICpercellpg, data=PIC)
E_DS_V_reg <- lm(E_DS_V~PICpercellpg, data=PIC)
plot(residuals.lm(E_DS_HV_reg))

plot(residuals.lm(E_DS_V_reg))
coef(E_DS_HV_reg)

 (Intercept) PICpercellpg 
    1.639233     3.243054 

coef(E_DS_V_reg)

 (Intercept) PICpercellpg 
 0.001639233  0.003243054 

#b. for liths
perlith_reg <- lm (perlithpg~PICpercellpg, data=PIC)
plot(resid(perlith_reg))

coef(perlith_reg)

  (Intercept)  PICpercellpg 
-6.547738e-17  5.000000e-02 

sinkvel_lith_reg <- lm(SinkVel_lith~PICpercellpg, data = PIC)
summary(sinkvel_lith_reg)

essentially perfect fit: summary may be unreliable

Call:
lm(formula = SinkVel_lith ~ PICpercellpg, data = PIC)

Residuals:
       Min         1Q     Median         3Q        Max 
-1.043e-16 -3.686e-17 -2.422e-18  4.040e-17  7.364e-17 

Coefficients:
              Estimate Std. Error   t value Pr(>|t|)    
(Intercept)  1.673e-02  8.512e-18 1.965e+15   <2e-16 ***
PICpercellpg 1.115e-02  1.109e-18 1.005e+16   <2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 4.579e-17 on 44 degrees of freedom
Multiple R-squared:      1, Adjusted R-squared:      1 
F-statistic: 1.011e+32 on 1 and 44 DF,  p-value: < 2.2e-16

plot(residuals.lm(sinkvel_lith_reg))
layout(matrix(1:4,2,2))

plot(sinkvel_lith_reg)

coef(sinkvel_lith_reg)

 (Intercept) PICpercellpg 
  0.01672753   0.01115169 

beta_lith_reg <- lm(beta_d_lith~PICpercellpg, data=PIC)
plot(residuals.lm(beta_lith_reg))
coef(beta_lith_reg)

 (Intercept) PICpercellpg 
2.302277e-07 1.528542e-07 

Elith_DS_HV_reg <- lm(Elith_DS_HV~PICpercellpg, data=PIC)
Elith_DS_V_reg <- lm(Elith_DS_V~PICpercellpg, data=PIC)
plot(residuals.lm(Elith_DS_HV_reg))

plot(residuals.lm(Elith_DS_V_reg))

coef(Elith_DS_HV_reg)

 (Intercept) PICpercellpg 
    2.302277     1.528542 

coef(Elith_DS_V_reg)

 (Intercept) PICpercellpg 
 0.002302277  0.001528542 

# 9. make new dataframe depending on experimental PIC values
# make a prediction based on PIC values
require(truncnorm)
require(Rmisc)
summary(PIC$PICpercellpg)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
-1.5873  0.5523  2.3009  4.6739  6.1204 20.1442 

summarySE(data=PIC, measurevar="PICpercellpg")

PIC_newdata <- as.data.frame(rtruncnorm(n=1000, a=-1.6, b=20.14, mean=4.7, sd=6.15))
#rename column. rename function in plyr 
library(plyr)
PIC_newdata <- rename (PIC_newdata, c ("rtruncnorm(n = 1000, a = -1.6, b = 20.14, mean = 4.7, sd = 6.15)" = 
                                         "PICpercellpg"))
PIC_newdata <- mutate(PIC_newdata, group = ifelse(PICpercellpg < 4 , "naked", "calcified"))
ggplotly(ggplot(data=PIC_newdata, aes(x=group, y=PICpercellpg)) + geom_boxplot()+geom_point(size=2) +
           theme_Publication())

PIC_newdata$group2 <- case_when(
  PIC_newdata$PICpercellpg <2  ~ "naked_bouyant",
  PIC_newdata$PICpercellpg >2 & PIC_newdata$PICpercellpg < 4 ~ "naked/calcified uncertain",
  PIC_newdata$PICpercellpg >4 & PIC_newdata$PICpercellpg < 10 ~ "moderately calcified",
  PIC_newdata$PICpercellpg >10 ~ "strongly calcified", 
  TRUE ~ as.character(PIC_newdata$PICpercellpg)
)
ggplotly(ggplot(data=PIC_newdata, aes(x=group2, y=PICpercellpg)) + 
           geom_boxplot()+geom_point(size=2) +theme_Publication())

#a. for host
PIC_newdata$SinkVel.pred <- predict(PIC_reg, data.frame(PIC_newdata))
PIC_newdata_reg <- lm(SinkVel.pred~PICpercellpg, data=PIC_newdata) 
coef(PIC_newdata_reg)

 (Intercept) PICpercellpg 
  0.01800852   0.01774764 

#same coef as PIC_reg
#> coef(PIC_newdata_reg)
#(Intercept) PICpercellpg 
#0.01800852   0.01774764 
plot(resid(PIC_newdata_reg))

ggplot(data=PIC_newdata, aes(x=PICpercellpg, y=SinkVel.pred)) +geom_point(size=2) +theme_Publication()+
  labs(y = expression("Predicted Sinking velocity"~("m"~day^-1)), x = expression("PIC"~cell^-1)) 

PIC_newdata$beta.pred <- predict(beta_reg, data.frame(PIC_newdata))
PIC_newdata$E_DS_V.pred <- predict(E_DS_V_reg, data.frame(PIC_newdata))
PIC_newdata$E_DS_HV.pred <- predict(E_DS_HV_reg, data.frame(PIC_newdata))
ggplot(data=PIC_newdata, aes(x=PICpercellpg, y=E_DS_V.pred)) +geom_point(size=5, aes(color=group2)) +
  theme_Publication() + geom_smooth() +
  scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) + annotation_logticks(sides="l")

  
PICbeta_new <- ggplot(data=PIC_newdata, aes(x=PICpercellpg, y=beta.pred)) +
  geom_point(size=5, aes(color=PICpercellpg))+
  scale_colour_gradient(name="PIC", guide=guide_colorbar(direction = "vertical", barheight=10))+
  theme_Publication() + 
  scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) + annotation_logticks(sides="l") +
  labs(y = expression(beta~("Predicted Encounters"~cm^3~day^-1)), x = expression("PIC"~cell^-1))+
  theme(legend.position = "right")
PICbeta_new

#b. for liths
PIC_newdata$perlithpg.pred <- predict(perlith_reg, data.frame(PIC_newdata))
PIC_newdata$SinkVel.pred.lith <- predict(sinkvel_lith_reg, data.frame(PIC_newdata))
sinkvel_lith_reg.pred <- lm(SinkVel.pred.lith~PICpercellpg, data=PIC_newdata) 
coef(sinkvel_lith_reg.pred)

 (Intercept) PICpercellpg 
  0.01672753   0.01115169 

#same coef as sinkvel_lith_reg
#> coef(sinkvel_lith_reg.pred)
#(Intercept) PICpercellpg 
# 0.01672753   0.01115169 
plot(resid(sinkvel_lith_reg.pred))

ggplot(data=PIC_newdata, aes(x=perlithpg.pred, y=SinkVel.pred.lith)) +geom_point(size=2) +theme_Publication()+
  labs(y = expression("Predicted Sinking velocity of Liths"~("m"~day^-1)), x = expression("PIC"~lith^-1)) 

PIC_newdata$beta.pred.lith <- predict(beta_lith_reg, data.frame(PIC_newdata))
PIC_newdata$E_DS_V.pred.lith <- predict(Elith_DS_HV_reg, data.frame(PIC_newdata))
PIC_newdata$E_DS_HV.pred.lith <- predict(Elith_DS_HV_reg, data.frame(PIC_newdata))
ggplot(data=PIC_newdata, aes(x=perlithpg.pred, y=E_DS_V.pred.lith)) +geom_point(size=5, aes(color=group2)) +
  theme_Publication() + scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) + annotation_logticks(sides="l")+ geom_smooth()

PICbeta_new.lith <- ggplot(data=PIC_newdata, aes(x=perlithpg.pred, y=beta.pred.lith)) +
  geom_point(size=5, aes(color=PICpercellpg))+
  scale_colour_gradient(name="PIC", guide=guide_colorbar(direction = "vertical", barheight=10))+
  theme_Publication() + 
  scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) + annotation_logticks(sides="l") +
  labs(y = expression(beta~("Predicted Encounters of lith " ~cm^3~day^-1)), x = expression("PIC"~lith^-1))+
  theme(legend.position = "right")
PICbeta_new.lith

#summaries
summary_DS_newdata_bygroup2 <- ddply(PIC_newdata, .(group2), summarize,  PICpercellpg=mean(PICpercellpg), SinkVel.pred=mean(SinkVel.pred),beta.pred= mean (beta.pred), E_DS_V.pred= mean(E_DS_V.pred), E_DS_HV.pred=mean(E_DS_HV.pred))
summary_DS_bygroup.pred <- ddply(PIC_newdata, .(group2), summarize,  PICpercellpg=mean(PICpercellpg), 
                            perlithpg.pred = mean(perlithpg.pred), SinkVel.pred=mean(SinkVel.pred),
                            beta.pred= mean (beta.pred), E_DS_V.pred= mean(E_DS_V.pred), 
                            E_DS_HV.pred=mean(E_DS_HV.pred), 
                            SinkVel.pred.lith=mean(SinkVel.pred.lith),beta.pred.lith= mean (beta.pred.lith),
                            E_DS_V.pred.lith= mean(E_DS_V.pred.lith), E_DS_HV.pred.lith=mean(E_DS_HV.pred.lith))
summary_DS_bygroup.pred

write.xlsx(summary_DS_bygroup.pred, file = "Postdoc-R/Exported Tables/summary_DS_bygroup.pred.xlsx")

cannot create file 'Postdoc-R/Exported Tables/summary_DS_bygroup.pred.xlsx', reason 'No such file or directory'

turbulence

#TURBULENCE
#disrate is cm2/s3
#make data frame
disrate <- rep_len(10^(-8:-2), length.out=14)
calc <- rep_len(c("calcified"), length.out=7)
naked <- rep_len(c("naked"), length.out=7)
group <- c(calc, naked)
turb <- as.data.frame(cbind(disrate, group))
turb <- mutate(turb, rad = ifelse(group == "naked" ,  1.8E-6,  2.3E-6)) #in m
turb$disrate <- as.numeric(as.character(turb$disrate))
turb$Kol <- ((v^3/turb$disrate)^0.25)*100 #Kolmogorov length scale in cm
#everything is below 1 cm, use eqn 2 in TK 
turb$beta_d <- (4.2*pi*((turb$disrate/(v*100^2))^0.5)*(((turb$rad+Rehv)*100)^3))*86400 
turb$beta_Heidi <- (0.42*pi*((turb$disrate/(v*100^2))^0.5)*(((turb$rad+Rehv)*100)^3))*86400
#check encounters
#use TK, in cm3 s
turb$E_turb_HV <- (turb$beta_d*hostnum*virnum) #E calculated with Virus and Host (10:1 MOI)
turb$E_turb_V <- (turb$beta_d*virnum) #E calculated with virus only
breaks <- 10^(-10:10)
minor_breaks <- rep(1:9, 21)*(10^rep(-10:10, each=9))
ggplot(data = turb, aes(x = disrate, y = beta_d, color=group)) + geom_point(size =5) +
  scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  scale_x_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  annotation_logticks() +
  theme_Publication()

library(scales)
ggplot(data = turb, aes(x = disrate, y = E_turb_V, color=group)) + geom_point(size =5) +
   scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  scale_x_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  annotation_logticks()+
  theme_Publication()

add beta kernels and plot

#extract mean betas from PIC_newdata
beta_DS <- summarySE (PIC_newdata, measurevar = "beta.pred", groupvars = c("group", "group2"))
all <- Reduce(function(x,y) merge(x,y,by="group",all=TRUE) ,list(BM, beta_DS, turb))
#beta_d.x=BM, beta.pred=DS, beta_d.y=turb
#rename beta.pred to beta_pred so I can use grep. 
#all <- rename (all, c("beta_d.x" = "beta_BM", "beta.pred" = "beta_DS", "beta_d.y" = "beta_turb"))
library(data.table)
NT = data.table(all, key="group2")
allbetas = NT[, list(group=group, disrate=disrate, beta_BM=beta_d.x, beta_DS=beta.pred, beta_turb = beta_d.y, 
                     beta_BM_DS =beta_d.x + beta.pred,
                     beta_DS_turb = beta.pred + beta_d.y,
                     beta_BM_turb = beta_d.x + beta_d.y,
                     beta_all = beta_d.x + beta_d.y + beta.pred), 
              by=c("group2")]
ggplot(allbetas, aes(disrate, y = value, color=group2)) + 
  geom_line(aes(y = beta_DS_turb, linetype = "DS+turb"), size=1) + 
  geom_line(aes(y = beta_all, linetype = "BM+DS+turb"), size=1)+
  geom_line(aes(y = beta_turb, linetype = "turb"), size=1)+
  scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  scale_x_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=7),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  annotation_logticks()+
  theme_Publication()

#encounters
#melt data
allbetas.melt <- melt (allbetas, id.vars = c("group2", "group", "disrate"), value.name = "beta_d", 
                       variable.name = "betakernel")
allbetas.melt$group2 <- factor (allbetas.melt$group2,levels= c("naked_bouyant", "naked/calcified uncertain",
                                                       "moderately calcified", "strongly calcified"),
                                                       labels = c("naked", "naked/calcified uncertain",
                                                                  "moderately calcified", "strongly calcified"))
allbetas.melt$E_V <- allbetas.melt$beta_d*virnum
ggplot(allbetas.melt, aes(disrate, y = E_V, color=group2)) + 
  geom_line(size=1)+
  scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  scale_x_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=7),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  annotation_logticks() + facet_grid(~betakernel)

#subset data
graph1 <- subset(allbetas.melt, betakernel %in% c ("beta_BM", "beta_DS", "beta_BM_DS"))
graph2 <- subset(allbetas.melt, betakernel %in% c ("beta_turb", "beta_DS_turb", "beta_BM_turb", "beta_all"))
ggplot(graph1, aes(group2, y = E_V, color=betakernel)) + 
  geom_jitter(size=5)+
  scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  annotation_logticks(sides = "l")+
  theme_Publication()

graph1.sum <- summarySE (graph1, measurevar = "E_V", groupvars = c("betakernel", "group2"))
ggplot(graph1.sum, aes(group2, y = E_V, color=betakernel)) + 
  geom_point(size=5,  position=position_dodge(0.2))+
  scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  annotation_logticks(sides = "l")+
  theme_Publication()

ggplot(data=allbetas.melt %>% filter(betakernel %in% c("beta_turb", "beta_all")), 
       aes(x=disrate,y = E_V, color=group2, linetype=betakernel)) + 
  geom_line(size=1, position=position_jitter(w=0.02, h=0))+
   scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=4),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  scale_x_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=7),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  annotation_logticks()+
  theme_Publication()

ggplot(data=allbetas.melt %>% filter(betakernel %in% c("beta_all")), 
       aes(x=disrate,y = E_V, color=group2)) + 
  geom_line(size=1, position=position_jitter(w=0.02, h=0))+
   scale_y_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=2),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  scale_x_log10(
        breaks = scales::trans_breaks("log10", function(x) 10^x, n=7),
        labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  annotation_logticks()+
  theme_Publication()